[Premium-Rx] RE: Receiver Frequency Display Accuracy & Calibration

[Alberto di Bene]

Peter Gottlieb <nerd at verizon.net> wrote:

> I just tried the PC spectrum analyzer (Spectran).  There definitely 
> seemed to be an offset in there as the 500 Hz WWV tone was showing up 
> around 512 Hz or so.  However, it did show that LSB and USB were 
> several Hz apart so I rechecked the master osc in my RF-350K and it 
> was indeed off a bit.  It was high about 0.04 ppm.  I tweaked it to be 
> right on (an academic proposition I am aware) and while Spectran still 
> showed an offset at least now both sidebands matched.
>
The precision of Spectran (and for that matter of any other software 
spectrum analyzer) is directly tied to the precision
of the sampling frequency of the sound card. Good quality sound cards 
have crystal oscillators for the various sampling
frequencies. Cheap cards do sample at the fixed rate of 48 kHz (and 
there maybe they are somewhat accurate), but for
the lower values the card driver does a resampling, which not always is 
much accurate, especially for non integral ratios
(e.g. 48 kHz subsampled at 44.1 kHz).  Accurary in subsampling costs CPU 
cycles, which could then cause a bad ranking
of the card in the benchmarks done by the various magazines. And, after 
all, who needs a precise sampling to playback
the sound of phasors, guns, explosions, laser torpedos and the likes ? 
(in case you are wondering if I have ever played a
PC game, the answer is a definite NO... :-)

73  Alberto  I2PHD